• 7 min read
How to Label Image Data for Computer Vision Models
This guide discusses what image labeling is and how to effectively label images for use in training computer vision models.
• 6 min read
You Might Be Resizing Your Images Incorrectly
Resizing images incorrectly can distort objects and degrade model performance. Learn how to resize images without distortion: aspect ratios, padding strategies, and practical tips to ensure your models learn from clean, consistent visuals.
• 11 min read
How to Grow Small Vision Datasets with SAM 3 and Synthetic Augmentation
Traditional image augmentation can make datasets bigger without making models better. In this tutorial, we show how to use SAM 3 to generate clean instance segmentation masks, apply segmentation-aware augmentation in Roboflow, and measure whether those changes actually improve model performance.
• 10 min read
How to Convert Annotations from PASCAL VOC XML to COCO JSON
Convert from VOC XML to COCO JSON (or any format) in four clicks.
• 3 min read
When to Use Grayscale as a Preprocessing Step
Grayscale allows our models to be more computationally efficient. So when **shouldn't** we grayscale our images?
• 9 min read
What Is Image Preprocessing and Augmentation?
Understanding image preprocessing and augmentation options is essential to making the most of your training data.
• 12 min read
Data Annotation for High-Performing Computer Vision Models
Learn all about data annotation, from what it is and how it works, to common challenges, best practices, and the tools that can streamline the process.
• 8 min read
5 Strategies for Handling Unbalanced Classes in Machine Learning
SUMMARY
Class imbalance causes models to underperform on minority classes because the training distribution does not reflect the real-world distribution of what matters. Five strategies address this: gathering more data from underrepresented classes, generating synthetic samples through augmentation, undersampling the majority class via random selection, oversampling the minority class
• 6 min read
Rock, Paper, Scissors with AI: How to Make Multiplayer Games
SUMMARY
HandLand is an open-source template for building two-player, gesture-controlled games in the browser using computer vision models from Roboflow Universe. The repo ships three complete games, including Rock Paper Scissors and a standoff game called 007, using WebRTC for peer video and native WebSockets for real-
• 3 min read
Launch: Find Similar Images to Expand Vision Datasets
Learn how to find similar images to use in your new computer vision datasets.
• 5 min read
Count Objects on a Conveyor Belt Using Computer Vision
SUMMARY
This tutorial builds an end-to-end system for counting bolts and nuts on a moving conveyor belt, covering every stage from data collection to local deployment. Video frames are annotated using Roboflow's auto-labeling with Grounding DINO, a custom detection model is trained in the cloud,
• 13 min read
How to Fine-tune PaliGemma 2
Learn how to fine-tune PaliGemma 2 to extract data from an image in JSON format.
• 11 min read
What is Active Learning? The Ultimate Guide.
In this guide, we discuss what active learning is, types of active learning, and walk through an example of active learning in practice.
• 6 min read
How to Create a Workout Pose Correction Tool
SUMMARY
This tutorial builds a real-time workout form checker using a custom Roboflow keypoint detection model trained to locate the left and right sides of a weight, combined with MediaPipe body tracking and deployed through Roboflow Workflows. The pipeline processes each video frame, extracts keypoints from both sources, and
• 3 min read
How to Import Hugging Face Datasets to Roboflow
Learn how to import a Hugging Face dataset into Roboflow for labeling, training, and deployment.
• 8 min read
What is the Open Images Dataset? A Deep Dive.
SUMMARY
Open Images V7, released by Google in 2022, contains over nine million annotated images spanning nearly 20,000 categories and supports six annotation types: image-level labels, bounding boxes, segmentation masks, visual relationships, localized narratives, and point-level labels. That breadth makes it one of the few public datasets
• 11 min read
How to Train RT-DETR on a Custom Dataset with Transformers
SUMMARY
RT-DETR (Real-Time DEtection TRansformer), developed by Peking University and Baidu, is a transformer-based object detection model that targets competitive speed and accuracy compared to YOLO-family models. This tutorial covers fine-tuning RT-DETR on a custom dataset sourced from Roboflow Universe, using the HuggingFace Transformers
• 4 min read
Import Images from Databricks to Roboflow
Upload image data from Databricks SQL warehouse into Roboflow for training custom computer vision models.
• 5 min read
How to Use YOLO-World With Active Learning to Train a Custom Model
In this guide, we demonstrate an approach where we can start using the benefits of YOLO-World now, while simultaneously collecting data to train a faster custom model later.
• 8 min read
Build Enterprise Datasets with CLIP for Multimodal Model Training Using Intel Gaudi2 HPUs
In this guide, learn how to use CLIP on Intel Gaudi2 HPUs to deduplicate datasets before training large multimodal vision models.
• 5 min read
How to Use Multiple Models to Label Datasets with Autodistill
In this guide, we cover the benefits of and how to combine multiple models in order to automatically label a dataset of images.
• 7 min read
Label Verification AI: How to Verify Label Placement on Packages
Learn how to build a system to verify label placement on packages using computer vision.
• 7 min read
How to Analyze a Folder of Videos from Google Cloud Platform
In this guide, we walk through how to analyze videos stored in Google Cloud Storage with computer vision models.
• 5 min read
Launch: Model Prompting for Automated Labeling with Autodistill
In this guide, learn how to use the Roboflow automated image labeling feature to label images in your computer vision datasets.
Stay Connected